EPR studies of cellular oxygen and nitroxide metabolism often involve a variety of experimental conditions including different sample preparations, cellular concentrations, and spectrometer settings. Post-processing of the resulting spectra also varies greatly. Here, we show that experimental conditions can affect results often in unexpected ways. We present data on the effects of cell density, type of cell bathing media, and nitroxide concentration in the sample, and power levels and modulation intensity setting of the spectrometer, on oxygen and nitroxide consumption results. We also compare information obtained from simple line height measurements with information obtained from least-squares spectral simulation techniques. Our most consistent results are obtained at low (105 cells/ml or less) cell density, media which alone does not reduce nitroxides, low (0.1mW) power, modulation levels appropriate for good (100:1 or better) signal-to-noise, and lineshape analysis by a fast convolution algorithm combined with Levenberg-Marquard optimization (for determining oxygen and nitroxide concentrations).